Method and apparatus for fast battery charging using neural network fuzzy logic based control
A method and apparatus for providing fast charging of secondary cells in an electronic device. The charging process is under the control of a microcontroller which contains a read-only-memory (ROM) in which is embedded code which determines the charging method. The charge method controls the charge provided to a battery back by a variable current source. An intelligent control scheme based on a neural network fuzzy logic methodology is used to optimize the charging current in response to measured characteristics of the battery.
Latest National Semiconductor Corporation Patents:
Claims
1. A method of fast charging a battery, comprising:
- providing a variable current source to supply current to charge the battery, wherein the current source delivers current at a plurality of levels;
- measuring a parameter indicative of an internal state of the battery; and
- controlling the level of charge supplied to the battery by the current source by means of a controller having an input and an output, wherein the input is a value of the measured parameter and the output of the controller is a signal which determines the level of current supplied by the current source, and further, wherein the output of the controller is determined according to the steps of
- inputting data representing a charge rate of the battery as a function of the measured parameter to a neural network;
- generating a fuzzy logic rule and membership function representing the battery's charging characteristics as an output of the neural network; and
- generating a set of instructions which control the operation of the controller to produce the controller output from the controller input, the instructions being generated from the fuzzy logic rule and membership function.
2. The method of claim 1, wherein the measured parameter is a member of the group consisting of the battery voltage, the battery temperature, and the current flowing into the battery.
3. The method of claim 1, wherein the generation of the fuzzy logic membership function and fuzzy logic rule by the neural network comprises the steps of:
- receiving a signal representing the input data;
- transforming the received signal into fuzzified data;
- inputting the fuzzified data into a neuron layer which generates an output which includes a signal corresponding to a fuzzy logic membership function; and
- inputting the membership signal into a neuron layer which generates an output which includes a logic rule signal representing a fuzzy logic rule.
4. A fast charging device for a battery, comprising:
- a variable current source to supply current to charge the battery, wherein the current source delivers current at a plurality of levels;
- a sensor for measuring a parameter indicative of an internal state of the battery; and
- a controller having an input and an output which controls the level of charge supplied to the battery by the current source, wherein the input to the controller is a value of the measured parameter and the output of the controller is a signal which determines the level of current supplied by the current source, and further, wherein the output of the controller is determined by executing a set of instructions which control the operation of the controller, the instructions being derived from a fuzzy logic membership function and fuzzy logic rule which represent the battery's charging characteristics, the controller further comprising
- a neural network having a charge rate of the battery as a function of the measured parameter as an input and the fuzzy logic membership function and fuzzy logic rule as an output.
5. The fast charging device of claim 4, wherein the sensor measures a parameter which is a member of the group consisting of the battery voltage, the battery temperature, and the current flowing into the battery.
6. The fast charging device of claim 5, wherein the neural network comprises:
- a first plurality of neurons for receiving a signal representing the input data for the controller and for providing fuzzified data which corresponds to the input data;
- a second plurality of neurons coupled to the first plurality of neurons for receiving the fuzzified data and generating a membership signal which corresponds to a fuzzy logic membership function; and
- a third plurality of neurons coupled to the second plurality of neurons for receiving the membership signal and generating a logic rule signal which represents a fuzzy logic rule.
| 5-244729 | September 1993 | JPX |
| 6-245403 | September 1994 | JPX |
- Yung et al., "A MicroController-based Battery Charger Using Neural-Fuzzy Technology" IEE 1994, Hong Kong. E. Khan, "Neural Network Based Algorithms for Rule Evaluation & Defuzzification in Fuzzy Logic Design," of National Semiconductor Corporation, Embedded Systems Division, 8 pages in length. E. Khan & P. Venkatapuram, "Neufuz: Neural Network Based Fuzzy Logic Design Algorithms," of National Semiconductor Corporation, Embedded Systems Division, FUZ-IEEE93, Mar. 28-Apr. 1, 1993, San Francisco, CA, 8 pages in length. E. Khan & P. Venkatapuram, "Fuzzy Logic Design Algorithms Using Neural Net Based Learning," of National Semiconductor Corporation, Embedded Systems Division, Presented at the Embedded Systems Conference, Santa Clara, Sep. 1992, 7 pages in length. J.A. Schofield, "Batteries keep up with portable electronics," Design News, Oct. 24, 1994, pp. 25-26. G. Cummings, D. Brotto & J. Goodhart, "Charge batteries safely in 15 minutes by detecting voltage inflection points," EDN, Sep. 1, 1994, pp. 89-92 & 94.
Type: Grant
Filed: Sep 8, 1994
Date of Patent: Feb 3, 1998
Assignee: National Semiconductor Corporation (Santa Clara, CA)
Inventors: Dilip S (Sunnyvale, CA), M. Zafar Ullah (San Jose, CA)
Primary Examiner: Edward Tso
Law Firm: Limbach & Limbach L.L.P.
Application Number: 8/302,602
International Classification: H01M 1046;
